Table 3

Fully adjusted HRs (95% CI) for severe renal disease associated with SGLT2i initiation versus non-SGLT2i initiation in patients with type 2 diabetes on metformin

SGLT2i exposure	N	Person-years	Severe renal disease cases	IR per 10 000 person-years	Adjusted HR (95% CI)*	P value*
ITT
Non-initiator	43 852	90 437	1104	122.1	1.0 (ref)
SGLT2i initiator	12 957	30 062	142	47.2	0.55 (0.46 to 0.67)	<0.01
OT
Non-initiator	43 852	81 643	1076	131.8	1.0 (ref)
SGLT2i initiator	12 957	16 933	43	25.4	0.34 (0.25 to 0.47)	<0.01
AT
Non-use	46 522	85 758	1131	131.9	1.0 (ref)
Current use of SGLT2i	18 469	29 681	72	24.3	0.31 (0.24 to 0.40)	<0.01
Recent use of SGLT2i	9706	1565	10	63.9	0.70 (0.38 to 1.32)	0.27
Past use of SGLT2i	5963	3495	33	94.4	0.82 (0.58 to 1.17)	0.28

Note that while ITT and OT categories are mutually exclusive, individuals might contribute to multiple AT categories during follow-up. Therefore, the sum of individuals contributing to each AT category exceeds the total number of individuals in the initial study cohorts.
*Adjusted estimates obtained using a Cox proportional hazards regression model including all variables in online supplemental table 1.
AT, as-treated; IR, incidence rate; ITT, intention-to-treat; OT, on-treatment; SGLT2i, sodium-glucose co-transporter-2 inhibitor.